We are using Redisson version 3.6.1 on AWS environment with ElastiCache setup on Replicated mode with 2 nodes. Our main operations are against an LRU bounded RMapCache (https://github.com/redisson/redisson/wiki/7.-distributed-collections#714-lru-bounded-map) where we just get and put data (having set a maxSize of 500K records).
However, we experience a series of org.redisson.client.RedisTimeoutException with message Unable to send command: .... even under really low load. Moreover, it seems that this phenomenon happens mainly on the primary node of the AWS ElastiCache. On some of our services which communicate with Redis (via Redisson) this doesn't really seem to be a problem, although such timeouts occur too, but less frequently.
The Redis ElastiCache cluster seems to be under no stress at all. CPU is usually to 1-2% utilisation. We have decreased the number of retryInterval and retryAttempts in order to fail-fast on such cases and prevent our service's threads from remaining blocked for long time. However, under that low load (about 100 redisson req/s) someone would expect to face no such issues at all.
